The open archive for STFC research publications

Full Record Details

Persistent URL http://purl.org/net/epubs/work/12145789
Record Status Checked
Record Id 12145789
Title Comparing Parallel Performance for a Geometric Multigrid Solver Using Hybrid Parallelism, MPI Shared Memory and Multiple GPUs
Abstract Performance optimisation of memory or communication bound applications, such as multigrid solvers, requires continuous evaluation of data placement and concurrency as the number of cores per compute node increases and memory hierarchies become more complex. These aspects are even more important if accelerators (GPUs or Xeon Phi) are used. For multicore nodes the standard method proposed to reduce the communication overheads is to use OpenMP for parallelism at node level and MPI for inter-node communication. However, scaling OpenMP towards 30 threads in a non-uniform memory access node proves to be difficult in many applications. Recently, the MPI 3 standard has introduced the ability to use shared memory for MPI ranks on the same node in order to reduce MPI data traffic while preserving the more structured communication environment offered by MPI. This is a study on the strong scaling for the DL_MG multigrid solver across several hardware platforms. We present a comparative performance analysis of various versions of hybrid OpenMP-MPI parallelism and MPI parallelism enhanced with the shared memory introduced by the MPI 3 standard. Performance data for a GPU implementation is also presented.
Organisation STFC , SCI-COMP
Keywords MPI performance , poisson equation , CCP , parallel multigrid
Funding Information
Related Research Object(s):
Licence Information:
Language English (EN)
Type Details URI(s) Local file(s) Year
Presentation Presented at 13th Copper Mountain Conference on Iterative Methods, Copper Mountain, Colorado, USA, 7-11 Apr 2014. dl-mg-CopperMountain-PRESENTED.pdf 2014